Exposure : Understanding Color and Grey Luminosity

Started Jul 12, 2007 | Discussions
Denyerec Regular Member • Posts: 475
Re: Exposure : Understanding Color and Grey Luminosity

I think the point is that your meter is colourblind, and certain colours although perceived by the human eye as bright (Such as the red swatch in the lower right) to a "colourblind" or luminosity-sensitive viewer they are not so.

Look at the red swatch with the lumo value of 50-something. It looks nice a punchy and bright. Now look at the grey with the 50-ish value. Very dark.

The result of this, seemingly, is that you may well find although a scene looks bright to your eyes, the meter is going to overexpose it, as that "bright red" is low in luminosity so your meter drags the exposure.

I don't think the OP is a "solution" to anything, moreover an interesting peice of data that one can assimilate and incorporate into ones thinking. EG, "bear in mind when shooting red roses that your meter is going to overexpose them" is simply on thing you can glean from this.

I may be totally up the wrong tree, but that's what I made of it. Someone with more knowledge and understanding better jump in and correct me if I'm totally wrong!!

Karld70 Veteran Member • Posts: 3,553
Excellent information /nt/

no text

Dave Wolfs Contributing Member • Posts: 664
Re: Exposure : Understanding Color and Grey Luminosity

Very cool. Thanks for the explanation.

So the meter only reads luminosity values?

I mean, suppose a scene was predominantly this punchy red color. With an LV value of 50 something. Does this mean that the camera would try to drag this value to around 128?

Denyerec wrote:

I think the point is that your meter is colourblind, and certain
colours although perceived by the human eye as bright (Such as the
red swatch in the lower right) to a "colourblind" or
luminosity-sensitive viewer they are not so.

Look at the red swatch with the lumo value of 50-something. It looks
nice a punchy and bright. Now look at the grey with the 50-ish value.
Very dark.

The result of this, seemingly, is that you may well find although a
scene looks bright to your eyes, the meter is going to overexpose it,
as that "bright red" is low in luminosity so your meter drags the
exposure.

I don't think the OP is a "solution" to anything, moreover an
interesting peice of data that one can assimilate and incorporate
into ones thinking. EG, "bear in mind when shooting red roses that
your meter is going to overexpose them" is simply on thing you can
glean from this.

I may be totally up the wrong tree, but that's what I made of it.
Someone with more knowledge and understanding better jump in and
correct me if I'm totally wrong!!

mof Forum Member • Posts: 77
Re: Exposure : Understanding Color and Grey Luminosity

I like Ron's post (and yes I have his book and DVD).

BUT - don't all of Nikon's DSLR cameras have colour built in to the Matrix meter?

(Maybe the purists will attack me for using Matrix rather than spot or centre weighted!)

Zane Paxton Veteran Member • Posts: 6,947
Presentation of Information

MK914 wrote:

Looking at square H-15, for instance... I see that the RGB value is
255 but its grey value is 64. Since this is sort of close to the 77
Grey value does this mean I can use a red rose in daylight like a
grey card? Maybe stopping down 1/3 or so? Just trying to figure how
to read all this.

The middle gray (or zone 5 or 18% reflectance) that the light meter sees is a value of 128 not 77. The practical approach if you were shooting an all bright red scene would be to decrease the exposure by half (one stop or -1 EV) to get the luminosity down from 128 that the meter sees to the luminosity value of 64 that red has.

Ron's intent is very good and thoughtful. However; what is confusing is how the information is organized and presented. As far as I can tell the color squares are fairly randomly placed on the chart. There may be some sort of disciplined progression of RGB values, but most humans can't understand and then process information that way (one of the advantages in Lab Color Processing is that it is more comprehensible). Also to have colors organized according to the zone system's classic f/stop steps would be more directly and practically understandable to a phtographer.

If the concept is aimed at understanding the luminosity of colors, the chart would be far more understandable (read practically useful) if it was organized and presented according to luminosity values with the corresponding gray value at the top of each column progressively according to the zone system. I acknowledge Ron's goal of digesting all this and then practicing it until it becomes innate as Roman suggests, but until one gets there the information would be more useful if it were to be organized in a way that is easier to immediately understand.

Hope that helps!

Cheers,
--
Zane
http://www.pbase.com/devonshire
Nikon D2x
NAPP Member

'Life is not measured by the number of breaths we take, but by the number of moments it takes our breath away.” ~ Anonymous

 Zane Paxton's gear list:Zane Paxton's gear list
Nikon D7000 Nikon AF-S DX Nikkor 16-85mm f/3.5-5.6G ED VR Nikon AF-S Nikkor 17-35mm f/2.8D ED-IF Nikon AF-S Nikkor 70-200mm F4G ED VR +2 more
OP Ron Reznick Veteran Member • Posts: 8,972
Based on your example...

Of course, I'd need to see the scene, but let's make some assumptions. It's a bright, clear day and the sky actually goes to blue-white at the horizon (the blue channel fully saturates at the horizon to the eye). There is about 25% of the frame covered by the sky, and the sky is roughly 170 blue and roughly 100 RGB average at the top of the frame. The green you mentioned is a combination of grass and foliage, ranging from 40RGB in the shadows (about 2.5% of the frame) to 225G in the brighter areas of the grass (about 30% of the frame). OK?

Now, based on the facts stated above, we know that the blue and green channels are at more risk than the red, and the risky regions are the sky at the horizon and the brightest areas of the grass. Depending on the metering mode you use, your decision-process would be different.

Let's say you are using a scene-meter (Matrix or Evaluative). This sort of scene is one well-covered in the on-board database, and it's likely that you would need very little EV adjustment unless there was a bright subject in the field of view, esp. in the part of the frame where there are darker areas. In that case, you'd need to bias the meter with negative EV -- the amount would be dependent on how bright that subject is and how dark the background. Knowing luminosity levels, and the distances between them in EV, is useful in this sort of situation, even when using a scene-meter.

If you were using either centerweight or spot meters, it's even more important to be able to evaluate the luminosity in the metered area, because the meter will set that area to the calibration point (generally about 116, with a little deviation), and you have to know both how bright that subject actually is, and how far that is from the metering calibration point so that you can set the bias EV.

With all this said, there are two general ways to expose any shot: one is to expose for reality, such that the subject or large areas of a scene are set to the same levels they are in reality, based on your best evaluation of the scene. The other is to expose for maximum use of the data bit-depth of the sensor (i.e. expose to the right). Your decision on which of these general methods to use should also be based on an evaluation of the luminosity in the scene and a knowledge of how the camera works... for instance, when shooting a very dark scene in which you need clean, detailed results of a dark subject in a dark background, it's to your advantage to push that exposure to get the data up in a region where there are more data points -- this will give you more detail and better gradients as well as less noise. If you can evaluate the scene luminosity and know how far it is from there to the level of exposure you want to achieve, you can make the correct EV adjustment prior to taking the shot.

This subject is rather complex. I've just touched on it here in this message and the previous one in this thread. I hope I've done enough for you to see some of the advantages to learning something about color, brightness levels, and how your camera works.

Ron
--
Ron Reznick
http://digital-images.net
http://trapagon.com

OP Ron Reznick Veteran Member • Posts: 8,972
Actually, that is not a dumb question

Fully saturated red with no other colors present only registers on those 25% of the sensor wells that have a red filter over them. What it means, is that if you were looking at a composite histogram, the red channel would saturate (hit 255) when the histogram showed a level of 64 (assuming no other color in the frame, and that the rose was pure red 255,0,0).

64 is -1.7EV from 128. The typical meter calibration point is about 116. If you filled the spot sensor (or the frame when using the centerweighted meter) with that rose you'd set -1.3EV assuming that it really was 255,0,0. That would achieve pretty darned close to the correct exposure on the first shot. Then, check the histogram and adjust if necessary for a second shot.

Ron
--
Ron Reznick
http://digital-images.net
http://trapagon.com

mcd3 Regular Member • Posts: 225
Re: Exposure : Understanding Color and Grey Luminosity

Well, Nikon metering is supposedly a 2005 spot color meter (or something like that), but I have never found it to be that fantastic a meter. In fact, I would rather use a hand-held spot meter, but that's a pain when working with small formats. I also do not think that a meter reading off of a rose is going to give you anything other than an exposure that will result in a middle value. The visual luminosity of that rose and your desired result will determine how to adjust your exposure, if necessary.

There used to be something called a monochrome viewing filter that beginning black and white photographers used. Looking through it approximated what the scene looked like without being influenced by the colors in the scene. Basically, it let you see values. Experienced photographers developed this understanding over time.

Of course, there are other factors to take into consideration, like the color of the light falling on the subject, which can influence how the value records on panchromatic film.

Those with cameras that allow shooting in black and white will learn a great deal about luminance by shooting in monochrome.

LilKnytt Veteran Member • Posts: 7,213
Since I'm the one who kind of started this......

I wish to thank some people here. Order is just as it comes to me....

I wish to thank Ron Reznick for seeing my plea & responding to it & I wish to thank you for taking a look at my gallery & pointing out this to me to start with. It lead to me wanting to push this issue into a learning tool for me. I've just ordered your complete set. Still debating if I can implore upon you in regards to the 2nd edition eBook which Jim Fenton recommends I try to get my hands on. I look at these charts, I've also downloaded the "gift" versions off your web site & printed them on high quality photo paper etc. for better reference, & know that I need text etc to better comprehend them & applying it to my photography.

Then there's Jim Fenton for taking time out to try to help with comments & recommendations for how to improve.

Then there's Roman, who's sweet enough to want to work on my .NEF file to see what he can do.

Then I've just realized that my spam filter had taken e-mails from Iain (Mustrum Ridcully) & Birger Petterson which I've just found & will respond to hopefully today. To both of you Thank You for taking the time out to care enough to e-mail me with help. I'm honored & truly appreciate your effort.

Anyhow, I'm happy to have opened up this "issue" & hope that I'll get a better understanding for what the camera sees & how. This will help me improve & that's what I want to do.

I'm happy to see that so many others can be helped by this & again, to all those who tried to help in my original thread, and to all others - thank you for help, support & effort I truly appreciate it.

Lil

-- hide signature --

A very humble beginning of a gallery, showing my progression with help from caring friends especially on DPR, can be visited by friends & family at

http://lilknytt.zenfolio.com/

Remember - - - it's humble, very humble

Julia Borg Veteran Member • Posts: 7,280
sorry Ron,

This
is the grey value that the camera sensor sees, based on the 25% Red,
50% Green, 25% Blue (Bayer Filter) over the sensor.

this theory is totally wrong, and not even close sometimes.

-- hide signature --

Julia

Zane Paxton Veteran Member • Posts: 6,947
A question for Ron

Ron Reznick wrote:

Fully saturated red with no other colors present only registers on
those 25% of the sensor wells that have a red filter over them. What
it means, is that if you were looking at a composite histogram, the
red channel would saturate (hit 255) when the histogram showed a
level of 64 (assuming no other color in the frame, and that the rose
was pure red 255,0,0).

64 is -1.7EV from 128.

Ron, how are you translating those luminosity values to f/stops or EV? I think that is where many people are not able to understand and then apply what you are presenting (me included).

In the classic zone system based on light being logarithmic, each f/stop either doubles or halves the amount of light. So to get from Luminosity of 128 to 64 is exactly one stop right? Unless the luminosity scales don’t match the zone system f/stop system? Please explain.

The typical meter calibration point is about
116. If you filled the spot sensor (or the frame when using the
centerweighted meter) with that rose you'd set -1.3EV assuming that
it really was 255,0,0. That would achieve pretty darned close to the
correct exposure on the first shot. Then, check the histogram and
adjust if necessary for a second shot.

Thanks!

-- hide signature --

Zane
http://www.pbase.com/devonshire
Nikon D2x
NAPP Member

'Life is not measured by the number of breaths we take, but by the number of moments it takes our breath away.” ~ Anonymous

 Zane Paxton's gear list:Zane Paxton's gear list
Nikon D7000 Nikon AF-S DX Nikkor 16-85mm f/3.5-5.6G ED VR Nikon AF-S Nikkor 17-35mm f/2.8D ED-IF Nikon AF-S Nikkor 70-200mm F4G ED VR +2 more
miklar Regular Member • Posts: 327
Re: sorry Ron,

Would you be so kind and explain where, or why, Ron is in error?
Cheers,
--
Michael Klar

Julia Borg Veteran Member • Posts: 7,280
Re: Presentation of Information

The middle gray (or zone 5 or 18% reflectance) that the light meter
sees is a value of 128 not 77.

middle grey is not 18%, 18% is not 128, and lightmeter does not see RGB.

The practical approach if you were
shooting an all bright red scene would be to decrease the exposure by
half (one stop or -1 EV)

and to get noise? not practical at all.

the practical approach is to realise that white balance affects histogram, while the exposure meter assumes UniWB.

-- hide signature --

Julia

Julia Borg Veteran Member • Posts: 7,280
Re: sorry Ron,

miklar wrote:

Would you be so kind and explain where, or why, Ron is in error?

he is in error with his theory of how luminosity values are calculated. I thought it was clear from my previous post.

-- hide signature --

Julia

miklar Regular Member • Posts: 327
Re: sorry Ron,

Julia Borg wrote:

miklar wrote:

Would you be so kind and explain where, or why, Ron is in error?

he is in error with his theory of how luminosity values are
calculated. I thought it was clear from my previous post.

Julia

I'm not being a contrarian, but am seriously interested in your understanding of the theory that does apply.

This is an issue I continuously try to grapple with and am open to listen and experiment with any and all theories and practial applications.
--
Michael Klar

Julia Borg Veteran Member • Posts: 7,280
Re: sorry Ron,

please read http://brucelindbloom.com/WorkingSpaceInfo.html

especially the "Side Notes" where Bruce addresses how to compute the luminance from RGB.

-- hide signature --

Julia

Zewt Regular Member • Posts: 437
Err, might have to agree with Julia. But I have no answer.

I have been working in LAB color for a few months now, and this posting only confuses me more.

miklar Regular Member • Posts: 327
Re: sorry Ron,

Julia Borg wrote:

please read http://brucelindbloom.com/WorkingSpaceInfo.html
especially the "Side Notes" where Bruce addresses how to compute the
luminance from RGB.

Thank you Julia

I will read (study) Burce's paper on the subject, admittedly, I have not heard of him before.
Cheers,
Michael
--
Michael Klar

Jeff Kohn Veteran Member • Posts: 4,855
Re: Presentation of Information

Julia Borg wrote:

the practical approach is to realise that white balance affects
histogram, while the exposure meter assumes UniWB.

I agree, if you really want to know what's going on in each channel, UniWB is the way to go.

One thing that bothered me in the exposure section of Ron's DVD is that he suggested changing the white balance to prevent individual channels from clipping. That seems like a great idea if you're just looking at the in-camera histogram, but it has absolutely no impact on the raw data stored in the NEF.

-- hide signature --
Dave Wolfs Contributing Member • Posts: 664
So are you basically saying to disregard the principles in this post?

Is what Ron saying completely wrong?

Keyboard shortcuts:
FForum MMy threads